IntenCheck API Guide
IntenCheck API is a lexicon-based sentiment engine which can provide a detailed analysis of texts, showing users what effects the words they use have on their readers. By using statistical methods the system extracts, identifies and characterizes the sentiment content from the text analyzed.
IntenCheck text analysis allows the user to:
- Evaluate the emotional characteristics of a text
- Detect personal characteristics of the writer
- Measure the effectiveness of the author’s communication by:
- comparing of the author’s intentions to impact;
- matching characteristics of a text to the communication style of the respondent(s).
IntenCheck monitors 7 different categories of content analyses within a text, which are arranged into 2 groups.
IntenCheck API detects 6 primary emotions: joy, surprise, anger, sadness, disgust, and fear so you know exactly how customers feel about your brand, campaign, and even your competition.
- Attitude (Semantic Differential)
Semantic differential is a type of a rating scale designed to measure the connotative meaning of objects, events, and concepts. IntenCheck API can determine the attitude of the author of the text toward the subject using 3 scales: positive-negative, active-passive, and strong-weak.
- Communication style
Refers to the four main ways in which we all process information. Some people tend to use more visual words, some use more auditory words and some tend to describe their feelings (kinesthetic). IntenCheck sentiment engine detects these 4 communication styles: visual, auditory, kinesthetic, and rational.
IntenCheck API checks if a text appears to be sincere or not, by comparing the number of words with vague or unspecific meaning to the general language norms or saved profiles.
Refers to the way a person thinks about time. IntenCheck detects past, present and future preferences within a text.
Refers to the way people are motivated either by moving away from something that they don’t desire or by moving towards to something that they want. IntenCheck shows the direction of motivation expressed in the text: away from or towards to.
- Perceptual positions
In communication, it’s important to understand how a person communicates from the inside out, either from his own viewpoint or from other perceptual positions. IntenCheck API detects 4 different perceptual positions: I, II, III, IV.
Getting started with the IntenCheck service
To use the free limited version of IntenCheck technology, please sign up for our service here which will allow you access to 4 categories of analyses (Emotions, Attitude, Communication Style, Perceptual Positions).
To sign up for IntenCheck API please contact us and we will provide you with further information on how to obtain API key.
For sentiment analysis of the text, IntenCheck engine uses language templates. A template is a generalized preference of communication that are common for a specific group of people, compiled from a massive language array. These are, for example, a group of people related professionally, demographically, or by some other characteristic and categorized by distinctive language attributes. Language templates can vary depending on the specific contexts – business, psychology, medical, scientific, etc.
IntenCheck API uses an American English language template as a baseline. This template was created by analyzing a significant amount of literature text used by the American High School curricula as well as periodicals.
By using language templates, IntenCheck engine calculates language norms for each group of categories. The language norm is a relative frequency of words from a specific category in the text.
In the future, there will be other specific language templates available, eg. business, economics, politics, etc. Users can also respond to a specific person by using the email(s) they received from their respondent and our engine will automatically compile a unique language norm which will be used in that case (“Answer to a person” feature). These unique language norms can also be saved as profiles and improved upon later, by adding more examples of texts or emails written by the same person.
Understanding the results
Intencheck API shows the following results:
wordsCount – the number of words corresponding to a particular category within the text.
totalWordsCount – is the total number of words in the text.
norm – is a number that represents the relative frequency of words from a specific category in the text. (For example, the relative frequency of words within a category Anger represent the average usage of the words for this category by the American English speakers.)
value – the likelihood (probability) that a certain characteristic of category can be attributed to the author/text.
- If the value < = 0.05, then the score is considered very low indicating that the characteristic of category can not be attributed to the author/text.
- If 0.05 < value < = 0.2, then the estimate is considered to be low and the characteristic of category cannot be attributed to the author/text.
- If 0.2 < value < 0.8 this means that the frequency of use of words in this category within the language norm range.
- If 0.8 <= value <0.95, then the estimate is considered high indicating the characteristic of a category that can be attributed to the author/text.
- If 0.95 < = value, then the estimate is considered to be very high, indicating that the characteristic of category can be attributed to the author/text.
words – a list of words from a specific category encountered in the text. These words are highlighted and used during editing.
Use of IntenCheck scale
For easier visualization, IntenCheck uses a scale from 0 to 100 (Value x 100). The scale is then broken down into 5 gradations and results are visualized using a thermometer scale for each gradation as per table below.
|0 – 5||Very Low||Contains very little statistical evidence. Very low probability to be perceived as a certain category.|
|6 – 20||Low||Below language norm. Low probability to be perceived as a certain category.|
|21 – 80||Normal||Within the language Norm.|
|81 – 95||High||Above language Norm. More likely to be perceived as a certain category.|
|96 – 100||Very High||Definitely statistically significant. Very high probability to be perceived as a certain category.|
Group of categories
Anger – evoked due to injustice, conflict, humiliation, negligence or betrayal. If anger is active, the individual attacks the target, verbally or physically. If anger is passive, the person silently sulks and feels tension and hostility.
Sadness – indicates a feeling of loss and disadvantage. When a person can be observed to be quiet, less energetic and withdrawn, it may be inferred that sadness exists.
Fear – a response to impending danger. It is a survival mechanism that is a reaction to some negative stimulus. It may be a mild caution or an extreme phobia.
Disgust – an emotional response of revulsion to something considered offensive or unpleasant. It is a sensation that refers to something revolting.
Surprise – evoked as the result of an unexpected event. Surprise can have any valence; that is, it can be neutral/moderate, pleasant, unpleasant, positive, or negative. Surprise can occur in varying levels of intensity ranging from very surprised, which may induce the fight-or-flight response or little surprise that elicits a less intense response to the stimuli.
Joy – joy or happiness has shades of enjoyment, satisfaction, and pleasure. There is a sense of well-being, inner peace, love, safety, and contentment.
The emotions results are displayed as per image below.
- Visual: language that creates pictures inside our mind;
- Auditory: language that creates sounds inside our mind;
- Kinesthetic: language that refers to the feelings that we feel inside our bodies;
- Rational: language that refers to logical thinking and thoughts.
Similar to Emotions measurement, Communication style is rated on 0 – 100 scale but visualized in a graphical manner (eg. below).
Insincerity is measured on 0-100 scale and displayed in a dial manner as per image below. The results from 80-100 show a high probability of insincere text.
Attitude (Semantic Differential)
Attitude scale results are displayed as a polarity graph and are graded as per table below. For each element of the pair, the scale is 0-100 and the value is displayed accordingly. The graph deviates towards the highest result for each pair.
The timeline scale reflects 3 subcategories: past, present, and future. Each tense is measured on a scale from 0-100 and results are displayed on a single graph as per image below.
Motivation scale shows two directions: Away From and Towards To. Each part is measured on a scale 0-100 and is displayed as per image below.
The perceptual position is graded on scale 0 – 100 and has 4 subcategories:
I – I, mine, self.
II – he, she, his, her, one, you, yours.
III – them, they, their.
IV – us, we, our.
The results are displayed as per image below.
Use IntenCheck engine to improve tone of the email
Below is an example where a sales manager is writing a reply email to a dissatisfied customer. It was analyzed using 4 categories from IntenCheck Free Online (limited) version.
The analysis showed that the initial email was Angry, Sad, Negative and too Rational which could lead to a permanent loss of a valued client.
It was then modified using the suggestions provided by the IntenCheck system and the final answer was much more Joyful and Positive, with the unifying perceptual Position-IV (we) dominating. This could change the situation with this customer to a positive outcome.
I am sorry to bother you with my emails.
I clearly see that you are displeased with the result, it’s quite alarming. I really want to find a suitable answer this problem as soon as possible.
Please excuse me and my colleague, for the previous email. It might sound a bit deceived, but I am sure he didn’t mean what he wrote, and he is very upset and frustrated about this outcome.
To resolve this issue, I would like to offer you free of charge service, for the first 6 months. After that, you will be able to decide whether to continue further collaboration or not.
Please do not hesitate to provide any feedback and thoughts on this proposal.
I am looking forward to your reply.
Enterprise SaaS Solutions Sales Manager
The results of Anger (96) and Sadness (85) represent the likelihood that these emotions are present within the text. Also, there is no statistical evidence that Joy (3) is present in the email.
There is strong statistical evidence that email is Negative (97).
Communication style results show that it’s slightly above Rational (86) norm.
Perceptual positions results for I (50) and IV (50) positions are within the language norm and positions II (26) and III (14) are less likely to be statistically significant.
It could be observed that the initial email is predominantly Angry, very Negative and also very Rational.
The email below was modified according to the suggestions provided by IntenCheck Online free version.
I hope you doing well.
We would like to thank you for your invaluable feedback and wish to resolve this situation as soon as possible.
Please excuse my colleague, for the previous email. I am sure he didn’t mean what he wrote to you, nevertheless we are optimistic about how we can turn it around to make you happy with our proposal.
To resolve this, I am delighted to offer you our services free of charge for the first 6 months. After that, we can discuss your experience with us and decide on any further collaboration.
Please do not hesitate to provide further feedback and your thoughts on this proposal. We strive to provide excellent services to our clients to have a wonderful experience.
I am looking forward to your reply and truly excited to work further with you.
Enterprise SaaS Solutions Sales Manager
The likelihood of Joy (80) present in the modified email is much more likely compared to the original version. The Anger (15) does not have any statistical significance, whereas Sadness (25) which is within the language norm and can be considered as a normal level.
There is a shift from Negative (42) to a more Positive (100) result which can be confirmed that there is strong evidence of positivity in the updated email.
Perceptual position IV [We] (100) now is the dominant one.
Summary for modified email
The modified email is more Joyful, Positive and is more balanced in terms of different communication styles with a domination of unifying (we) perceptual Position-IV.
Currently, only English text can be analyzed.
Questions and feedback
Find answers to your questions about our technology in our knowledge base here.
Please submit your questions, comments or feedback by filling a ticket form here or email us directly: email@example.com.
Science behind IntenCheck categories
Matching language patterns
One of the main features that our software engine offers to its users refers to an idea first introduced in the field of neuro-linguistic programming: that of matching language patterns, particularly communication style (representational systems).
By being aware of a person’s language and then using their preferred communication style, we can build rapport with them faster and get our message across more clearly. It also makes it more likely to have our point of view accepted with less resistance. Essentially, matching language patterns helps us better understand others and how they think and leads to a greater ability to influence.
The idea of matching elements of communication gives the user a clear idea about the direction of his/her communication.
Language templates and profiles
In order to be able to compare these elements of communication, our software uses language templates. These are generalized preferences of communication that are common for most people. Language templates can vary depending on the specific contexts – business, psychology, medical, scientific, etc.
Users can also respond to a specific person by using the email(s) they received from their respondent and our engine automatically compiles a unique language norm which will be used in that case. These unique language norms can also be saved as profiles and improved upon later, by adding more examples of texts or emails written by the same person.
Sentiment analysis methods and tools have been developed for a number of years, with one of the earliest works about this field being published in 2002 by Turney and Pang:
– Peter Turney. “Thumbs Up or Thumbs Down? Semantic Orientation Applied to Unsupervised Classification of Reviews”;
– Bo Pang; Lillian Lee and Shivakumar Vaithyanathan. “Thumbs up? Sentiment Classification using Machine Learning Techniques”;
Sentiment analysis aims to determine the attitude of a speaker or writer regarding a topic or the overall contextual polarity of a document. The attitude may be his or her judgment or evaluation, affective state (emotional state of the author when writing), or the intended emotional communication (the emotional effect the author wishes to have on the reader).
One of the things we noticed when we were researching our competition is that they all focus on analyzing and classifying the sentiment of a text in a single category of results and determining whether it is “Positive”, “Neutral” or “Negative”. With the exception of one competitor, who analyses emotions or the tone of a text, all the other companies that provide sentiment analysis rely on this classification.
Intentex Ltd. has developed a text analysis engine which is bringing a whole lot more to the sentiment analysis market. Our tools support 7 categories of results.
Paul Ekman, Ph.D. who has been at the forefront of researching and classifying emotions has concluded that there are six basic human emotions. Our sentiment engine can analyze a text and show how it scores for each of these. It can determine what kind of emotions it creates in the reader/listener. The six basic human emotions are:
Examples of existing research
When developing our system, we have looked also at the different studies and more complex classifications of emotions, some of the most important being Dr. Robert Plutchik ‘wheel of emotions’ (1980), where he demonstrated how different emotions can blend into one another and create new emotions, and the ‘classification of emotions’ by W. Gerrod Parrott, Ph. D, in which he identified over 100 emotions and conceptualized them as a tree-structured list.
Some of the studies that we have used when developing this category are:
- Ekman, P., Friesen, W. V., & Ellsworth, P. (1982). What emotion categories or dimensions can observers judge from facial behavior?
- Ekman, Paul (1992). “An argument for basic emotions”.
- Plutchik, Robert (1980), Emotion: Theory, research, and experience: Vol. 1. Theories of emotion 1, New York: Academic
- Plutchik, R. (2002). Nature of emotions. American Scientist
- Parrott, W. (2001), Emotions in Social Psychology, Psychology Press, Philadelphia
- Oatley, K., & Johnson-Laird, P. N. (1987). Towards a cognitive theory of emotions. Cognition & Emotion.
- Frijda, N.H., Mesquita, B., Sonnemans, J., & Van Goozen, S. (1991). The duration of affective phenomena or Emotions, sentiments, and passions. In K.T. Strongman (Ed.), International Review of Studies on Emotion, vol. 1. New York: Wiley.
- Izard, C. E. (1991). “The Psychology of Emotions”.
- Ortony, A., & Turner, T. J. (1990). What’s basic about basic emotions? Psychological Review
- Intentex White Paper about the ‘Emotions’ category: http://intentex.com/sentiment-analysis-emotions-white-paper/.
2. Communication Style
As part of the neuro-linguistic programming field (NLP), the communication style refers to the four main ways in which we all process information inside our mind. Our sentiment engine shows how this works and gives suggestions on how to match this style for the person you are communicating to so that your message is processed easier, faster and understood better.
– Visual: language that creates pictures inside our mind;
– Auditory: language that creates sounds inside our mind;
– Kinesthetic: language that refers to the feelings that we feel inside our bodies;
– Rational: language that refers to logical thinking and thoughts.
Examples of existing research
The category of results that we have named “Communication style” is referred to as “representational systems” within the field of NLP. Representational systems, also known as sensory modalities, a model from neuro-linguistic programming that examines how the human mind both takes in data and then processes and stores information.
More detailed information about representational systems can be found in published works such as:
– Bandler, Richard; Grinder, John (1976). The Structure of Magic II. Science and behavior Books Inc.
– Grinder, John & Carmen Bostic St Clair (2001.). Whispering in the Wind. CA: J & C Enterprises.
– Bandler, Richard & John Grinder (1979). Frogs into Princes: Neuro Linguistic Programming. Moab, UT: Real People Press.
– Dilts, Robert B, Grinder, John, Bandler, Richard & DeLozier, Judith A. (1980). Neuro-Linguistic Programming: Volume I – The Study of the Structure of Subjective Experience. Meta Publications, 1980.
Intentex White Paper about the ‘Communication Style’ category: http://intentex.com/sentiment-analysis-communication-style-white-papers/.
Our software checks if a text appears to be sincere or not, by comparing the number of words with vague or unspecific meaning to the general language norms or saved profiles.
When it comes to texts, it is very difficult to detect when somebody is lying, but when somebody is trying to cover or not give away all the truth, the language they use can signal their intention of withholding information.
Studies show that when we try to conceal our thoughts, we use more vague words than we would normally do. If there is a statistically significant increase in the use of such words within a text (when compared to the conventional language norm or to the profile that we have created for that person) it can show that the author is not telling something or trying to hide parts of information.
Examples of existing research:
In order to find a way of detecting signs of insincerity, we have looked at a number of different studies and reports, such as:
– Preston, Elizabeth (July 2002). “Detecting Deception”;
– Spence SA, et al. Behavioural and functional anatomical correlates of deception in humans. Neuroreport. (2001);
– The Truth About Lie Detectors. American Psychological Association.
Intentex White Paper about the ‘Insincerity’ category: http://intentex.com/sentiment-analysis-insincerity-white-papers/.
Not only do we classify texts as “positive” or “negative”, we use semantic differential scales to show the exact range and scores for the attitudes expressed. These includes the 3 main scales that all humans unconsciously use to interpret attitudes:
- Positive – Negative
- Strong – Weak
- Active – Passive
Examples of existing research
In order to measure the attitude expressed within a text, our system uses a semantic differential to categorize words based on their connotative meaning. A semantic differential scale is a list of opposite adjectives, and it is a method invented by Charles E. Osgood, Ph.D. in 1957 in order to measure the connotative meaning of cultural objects.
The three scales that our software uses were found to be the most relevant for measuring attitudes and they are cross-cultural universals in a study of dozens of cultures.
The semantic differential scale is today one of the most widely used in the measurement of attitudes.
More detailed information about it can be found in published works such as:
- Himmelfarb, S. (1993). The measurement of attitudes. In A.H. Eagly & S. Chaiken (Eds.), Psychology of Attitudes. Thomson/Wadsworth.
- Osgood, C.E., Suci, G., & Tannenbaum, P. (1957) The measurement of meaning. Urbana, IL: University of Illinois Press.
- Snider, J. G., and Osgood, C. E. (1969) Semantic Differential Technique: A Sourcebook. Chicago: Aldine.
Intentex White Paper about the ‘Attitude’ category: https://intentex.com/sentiment-analysis-attitude-white-papers/.
Timeline refers to the way a person thinks about time and it is another concept taken from neuro-linguistic programming. In this case, we refer to the basic ideas like ‘before’ and ‘after’, ‘past’, ‘future’, and ‘now’.
Some people tend to live more ‘in the moment’, others have a tendency to think more about the future and “things to come”, while others seem to be stuck in the past, constantly thinking about and reliving some of the events that have already happened. By understanding how texts communicate time concepts, users can greatly improve their results by matching the preferences of their reader/listener.
Examples of existing research:
The concept of the timeline is extensively studied in the field of neuro-linguistic programming and it has been further detailed in works such as:
– Hall, L. Michael & Bob G. Bodenhamer (1997), Figuring People Out: Design Engineering with Meta Programs, Crown House Publishing Ltd.
– Hall, L. Michael & Bob G. Bodenhamer (1997), Mind Lines: Lines for changing minds, E.T. Publications.
– Intentex’s white paper about the ‘Timeline’ category: http://intentex.com/sentiment-analysis-timeline-white-paper/.
6. Motivation direction
This refers to another part of the neuro-linguistic programming field and it shows the direction of motivation expressed in a text:
– Away from: refers to motivating an individual by moving away from unpleasant/unwanted things;
– Towards: refers to motivating an individual by moving towards what is desired or wanted.
Examples of existing research:
The direction of motivation is studied in the field of neuro-linguistic programming as one of the meta programs.
Meta programs are advanced patterns of thinking that control what we perceive, or in other words, they are mental processes which manage, guide and direct other mental processes.
The motivation direction is also studied extensively and considered an integral part of the “LAB Profile”. This system of profiling is used to predict behavior based on one’s use of certain types of language patterns (which reflect the thinking styles of a person and indicates their motivation and attitude).
A detailed explanation regarding the direction of motivation can be found in the following published works:
– Charvet, Shelle R. (1997). Words that change minds: Mastering the language of influence (2nd Revised edition). Hushion House.
– Hall, L. Michael & Bob G. Bodenhamer (2006), Figuring People Out: Reading People using Meta Programs, Crown House Publishing Ltd.
Intentex White Paper about the ‘Motivation Direction’ category: http://intentex.com/sentiment-analysis-motivation-white-papers/.
7. Perceptual Positions
The notion of perceptual positions is also one that is taught in the field of neuro-linguistic programming and it describes how the point of view that we adopt in any given situation can influence our perception.
IntenCheck’s results in this category offer invaluable information regarding the point of view expressed within a text, and can lead to many insights about the most effective ways to communicate:
– Position I: Thinking and communicating from one’s own experience and point of view (typical of leaders);
– Position II: Thinking and communicating while thinking about the other person’s experience (shows empathy and understanding);
– Position III: Thinking and communicating as an external observer, as if uninvolved (feedback appears as objective);
– Position IV: Thinking and communicating from the point of view of the entire system (gives the impression of an organization/company/group of people acting and thinking in the same way);
Examples of existing research:
The importance of perceptual positions (or points of view) has been detailed in works such as:
– Bodenhamer, Bob G. & L. Michael Hall. The User’s Manual for the Brain (Vol 1).
– Taylor, Shelley E. & Susan T. Fiske. Social Cognition.
– Lassiter, G. Daniel & Christian A. Meissner. Police Interrogations and False Confessions: Current Research, Practice, and Policy.
– Jago, Wendy & Ian McDermott. The NLP Coach: A comprehensive guide to personal well-being & professional success.
Intentex White Paper about the ‘Perceptual Positions’ category: http://intentex.com/sentiment-analysis-perceptual-positions-white-paper.